Robots are used extensively in different applications so as to perform specific tasks. However, the robots are required to move around in a location where they are present, so as to perform the tasks. For path planning, free space identification is performed by the robots during which obstacles are detected and free space is identified. However, the existing systems for path planning struggle to identify free space in cluttered environments. The disclosure herein generally relates to robotic path planning, and, more particularly, to a method and system for free space detection in a cluttered environment for robotic path planning. The system inscribes obstacles in bounding boxes and all unit grids inscribed by the bounding boxes are considered as occupied. Further, by seeding the occupancy grid map, the system identifies unified segments and corresponding seeds. Further a convex expansion is executed in the occupancy grid map to detect the free space.
Legal claims defining the scope of protection. Each claim is shown in both the original legal language and a plain English translation.
1. A processor-implemented method for free space detection for robotic path planning, comprising: fetching, via one or more hardware processors, sensor data pertaining to location of at least one obstacle in an area being scanned by a robot; generating an occupancy grid map of the area based on the sensor data, via the one or more hardware processors, wherein the at least one obstacle is inscribed in a bounding box, in the occupancy grid map; marking, via the one or more hardware processors, all unit geometric grids of the occupancy grid map which are fully or partially inscribed in the bounding box, as occupied; seeding the occupancy grid map, via the one or more hardware processors, comprising: generating a plurality of unified segments, wherein each of the plurality of unified segments is generated by merging the unit geometric grids which are out of the at least one bounding box; identifying center of each of the plurality of unified segments; and marking the identified center of each of the plurality of unified segments as a seed of the corresponding unified segment; and identifying the free space by executing a convex expansion in the occupancy grid map, via the one or more hardware processors, comprising: creating a convex region around each seed; iteratively expanding the convex region within the corresponding unified segment, in all directions, till the edges of the convex region touch either an edge of the unified segment or an edge of the obstacle; and identifying area covered by all the convex regions in the occupancy grid map as the free space.
This invention relates to robotic path planning, specifically a method for detecting free space in an environment using sensor data. The method involves processing sensor data to identify obstacles and determine navigable areas for a robot. The process begins by collecting sensor data about obstacle locations in a scanned area. An occupancy grid map is generated, where obstacles are enclosed within bounding boxes. All grid units fully or partially within these bounding boxes are marked as occupied. The method then seeds the map by merging adjacent unoccupied grid units into unified segments, identifying the center of each segment, and marking these centers as seeds. Free space is identified by expanding convex regions around each seed in all directions until the regions encounter obstacle edges or segment boundaries. The combined area of all convex regions defines the free space available for robotic navigation. This approach efficiently maps free space by leveraging geometric expansion from seeded points, improving path planning accuracy and computational efficiency.
2. The method as claimed in claim 1 , wherein the plurality of unified segments represent free regions in the occupancy grid map.
This invention relates to occupancy grid mapping, a technique used in robotics and autonomous navigation to represent the environment as a grid of cells indicating occupied or free space. The challenge addressed is efficiently identifying and utilizing free regions within the occupancy grid to optimize path planning, obstacle avoidance, or other navigation tasks. The method involves generating a plurality of unified segments from the occupancy grid map, where each segment represents a contiguous free region. These segments are derived by analyzing the grid to group adjacent free cells into larger, coherent areas. The unified segments can then be used to improve navigation by providing a more structured representation of traversable space, reducing computational overhead compared to processing individual grid cells. The method may include additional steps such as refining the segments based on geometric constraints, merging overlapping segments, or filtering segments below a certain size threshold to ensure relevance. The unified segments can be further processed to extract features like centroids, boundaries, or connectivity information, which are useful for path planning algorithms or dynamic obstacle avoidance. By representing free regions as unified segments rather than discrete cells, the method enhances efficiency in navigation systems, particularly in environments with complex or dynamic obstacles. This approach is applicable in autonomous vehicles, robotic systems, and other applications requiring real-time spatial awareness.
3. The method as claimed in claim 1 , wherein while expanding the convex region in a unified segment, other unified segments are marked as virtual obstacles.
A system and method for path planning in robotic navigation or autonomous vehicle control addresses the challenge of efficiently navigating through complex environments with dynamic obstacles. The invention involves expanding a convex region representing a safe navigation area while treating other predefined segments as virtual obstacles to ensure collision-free movement. The method dynamically adjusts the navigation path by expanding the convex region in a unified segment, which is a contiguous area defined by spatial constraints. During this expansion, other unified segments within the environment are marked as virtual obstacles to prevent collisions and ensure the path remains within safe boundaries. This approach enhances path optimization by considering both physical and virtual constraints, improving navigation efficiency in dynamic environments. The system may integrate sensor data, such as LiDAR or camera inputs, to update obstacle positions in real-time, ensuring adaptive path planning. The method is particularly useful in applications requiring precise movement, such as warehouse robots, autonomous drones, or self-driving cars, where avoiding collisions with both static and dynamic obstacles is critical. The invention provides a robust solution for real-time path planning by combining convex region expansion with virtual obstacle marking, ensuring safe and efficient navigation.
4. A system for free space detection for robotic path planning, comprising: a memory module storing a plurality of instructions; one or more communication interfaces; and one or more hardware processors coupled to the memory module via the one or more communication interfaces, wherein the one or more hardware processors are caused by the plurality of instructions to: fetch sensor data pertaining to location of at least one obstacle in an area being scanned by a robot; generate an occupancy grid map of the area based on the sensor data, wherein the at least one obstacle is inscribed in a bounding box in the occupancy grid map; mark all unit geometric grids of the occupancy grid map which are fully or partially inscribed in the bounding box, as occupied; seed the occupancy grid map, by: generating a plurality of unified segments, wherein each of the plurality of unified segments is generated by merging the unit geometric grids which are out of the at least one bounding box; identifying center of each of the plurality of unified segments; and marking the identified center of each of the plurality of unified segments as a seed of the corresponding unified segment; and identify the free space by executing a convex expansion in the occupancy grid map, by: creating a convex region around each seed; iteratively expanding the convex region within the corresponding unified segment, in all directions, till the edges of the convex region touch either an edge of the unified segment or an edge of the obstacle; and identifying area covered by all the convex regions in the occupancy grid map as the free space.
This system addresses the challenge of accurately detecting free space for robotic path planning in environments with obstacles. The system uses sensor data to identify obstacle locations and generates an occupancy grid map, where obstacles are enclosed within bounding boxes. All grid units fully or partially within these boxes are marked as occupied. The system then processes the map by merging adjacent unoccupied grid units into unified segments, identifying the center of each segment, and marking these centers as seeds. Free space is determined by expanding convex regions from each seed outward until they encounter obstacle edges or segment boundaries. The combined area of all convex regions defines the navigable free space. This approach improves path planning by efficiently identifying continuous free regions while accounting for obstacle boundaries, enhancing robotic navigation in dynamic environments. The system leverages hardware processors and memory modules to execute these steps, ensuring real-time processing for robotic applications.
5. The system as claimed in claim 4 , wherein the plurality of unified segments represent free regions in the occupancy grid map.
This invention relates to autonomous navigation systems that use occupancy grid maps to represent environments. The problem addressed is efficiently identifying and utilizing free regions in such maps for path planning and obstacle avoidance. The system generates a unified representation of free regions by segmenting the occupancy grid map into multiple unified segments. These segments are derived from the grid map data, where each segment corresponds to a distinct free region that is navigable by the autonomous system. The segmentation process involves analyzing the grid map to detect contiguous free cells and grouping them into coherent segments. These segments are then used to optimize path planning by reducing computational complexity and improving navigation efficiency. The unified segments can be dynamically updated as new sensor data is received, allowing the system to adapt to changing environments in real time. This approach enhances the accuracy and reliability of autonomous navigation by providing a structured representation of free space, enabling more efficient and safer movement through the environment.
6. The system as claimed in claim 4 , wherein said system is configured to mark all other unified segments as virtual obstacles while expanding the convex region in a unified segment.
This invention relates to path planning and navigation systems for autonomous vehicles or robotic systems, particularly in environments with dynamic or complex obstacles. The system addresses the challenge of efficiently expanding a convex region within a segmented environment while avoiding collisions with obstacles. The core functionality involves identifying and marking other unified segments as virtual obstacles during the expansion process to ensure safe and optimal path planning. A unified segment represents a contiguous area within the environment, and the convex region expansion is a method for determining feasible movement paths. By treating other segments as virtual obstacles, the system prevents collisions and ensures that the expanded region remains collision-free. This approach is particularly useful in dynamic environments where obstacles may change position or configuration over time, requiring real-time adjustments to the path planning process. The system dynamically updates the virtual obstacle designations as the convex region expands, allowing for adaptive navigation in complex scenarios. The overall goal is to improve the efficiency and reliability of autonomous navigation by incorporating obstacle avoidance directly into the path expansion process.
7. A non-transitory computer readable medium for free space detection for robotic path planning, comprising instructions, which when executed by one or more hardware processors causes the one or more hardware processors to perform a method comprising: fetching, via the one or more hardware processors, sensor data pertaining to location of at least one obstacle in an area being scanned by a robot; generating an occupancy grid map of the area based on the sensor data, via the one or more hardware processors, wherein the at least one obstacle is inscribed in a bounding box, in the occupancy grid map; marking, via the one or more hardware processors, all unit geometric grids of the occupancy grid map which are fully or partially inscribed in the bounding box, as occupied; seeding the occupancy grid map, via the one or more hardware processors, comprising: generating a plurality of unified segments, wherein each of the plurality of unified segments is generated by merging the unit geometric grids which are out of the at least one bounding box; identifying center of each of the plurality of unified segments; and marking the identified center of each of the plurality of unified segments as a seed of the corresponding unified segment; and identifying the free space by executing a convex expansion in the occupancy grid map, via the one or more hardware processors, comprising: creating a convex region around each seed; iteratively expanding the convex region within the corresponding unified segment, in all directions, till the edges of the convex region touch either an edge of the unified segment or an edge of the obstacle; and identifying area covered by all the convex regions in the occupancy grid map as the free space.
This invention relates to robotic path planning, specifically a method for detecting free space in an environment using sensor data. The system addresses the challenge of accurately identifying navigable areas for robots by processing sensor inputs to generate an occupancy grid map. The map represents the environment as a grid where obstacles are enclosed within bounding boxes. All grid units fully or partially within these boxes are marked as occupied. The method then processes the remaining unoccupied grid units by merging adjacent units into unified segments. For each segment, the center is identified and marked as a seed point. Convex expansion is applied around each seed, iteratively expanding in all directions until the convex region encounters either the segment boundary or an obstacle edge. The combined area of all convex regions is designated as free space, providing a clear representation of navigable areas for path planning. This approach improves upon traditional methods by efficiently segmenting and expanding regions to accurately determine free space, enhancing robotic navigation in dynamic environments. The system leverages computational geometry to optimize path planning by reducing the complexity of obstacle representation and free space identification.
8. The non-transitory computer readable medium as claimed in claim 7 , wherein the plurality of unified segments represent free regions in the occupancy grid map.
This invention relates to autonomous navigation systems that use occupancy grid maps to represent environments. The problem addressed is efficiently managing and updating these maps to accurately reflect free and occupied regions for navigation. The invention involves a non-transitory computer-readable medium storing instructions for generating and processing an occupancy grid map. The map is divided into a plurality of unified segments, where each segment represents a distinct region of the environment. These segments are used to identify free regions, allowing autonomous systems to plan paths through unobstructed areas. The unified segments are dynamically updated based on sensor data, ensuring the map remains accurate as the environment changes. The system may also include methods for merging or splitting segments to optimize map resolution and computational efficiency. By structuring the map into unified segments, the invention improves the accuracy and efficiency of path planning in autonomous navigation.
9. The non-transitory computer readable medium as claimed in claim 7 , wherein while expanding the convex region in a unified segment, other unified segments are marked as virtual obstacles.
A system and method for path planning in robotic navigation involves expanding a convex region within a navigation space to determine feasible paths while avoiding obstacles. The technology addresses the challenge of efficiently navigating complex environments by dynamically adjusting the navigation space to account for both physical and virtual obstacles. When expanding a convex region to explore potential paths, the system marks other unified segments as virtual obstacles to prevent collisions or conflicts during movement. Unified segments represent predefined path segments or regions that must be treated as impassable during the expansion of a particular convex region. This approach ensures that the path planning algorithm avoids overlapping or interfering with other segments, maintaining safe and collision-free navigation. The method is particularly useful in environments where multiple robots or moving objects require coordinated path planning to avoid conflicts. By treating certain segments as virtual obstacles, the system dynamically adapts to changing conditions, improving navigation efficiency and safety. The solution is implemented using a non-transitory computer-readable medium containing instructions for executing the path planning algorithm, ensuring robust and reliable operation in real-world applications.
Cooperative Patent Classification codes for this invention. Click any code to explore related patents in that topic.
September 20, 2019
March 8, 2022
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.